Simultaneous Decolorization and Biohydrogen Production from Xylose by Klebsiella oxytoca GS-4-08 in the Presence of Azo Dyes with Sulfonate and Carboxyl Groups

ABSTRACT

Biohydrogen production from the pulp and paper effluent containing rich lignocellulosic material could be achieved by the fermentation process. Xylose, an important hemicellulose hydrolysis product, is used less efficiently as a substrate for biohydrogen production. Moreover, azo dyes are usually added to fabricate anticounterfeiting paper, which further increases the complexity of wastewater. This study reports that xylose could serve as the sole carbon source for a pure culture of Klebsiella oxytoca GS-4-08 to achieve simultaneous decolorization and biohydrogen production. With 2 g liter⁻¹ of xylose as the substrate, a maximum xylose utilization rate (UR_xyl) and a hydrogen molar yield (HMY) of 93.99% and 0.259 mol of H₂ mol of xylose⁻¹, respectively, were obtained. Biohydrogen kinetics and electron equivalent (e⁻ equiv) balance calculations indicated that methyl red (MR) penetrates and intracellularly inhibits both the pentose phosphate pathway and pyruvate fermentation pathway, while methyl orange (MO) acted independently of the glycolysis and biohydrogen pathway. The data demonstrate that biohydrogen pathways in the presence of azo dyes with sulfonate and carboxyl groups were different, but the azo dyes could be completely reduced during the biohydrogen production period in the presence of MO or MR. The feasibility of hydrogen production from industrial pulp and paper effluent by the strain if the xylose is sufficient was also proved and was not affected by toxic substances which usually exist in such wastewater, except for chlorophenol. This study offers a promising energy-recycling strategy for treating pulp and paper wastewaters, especially for those containing azo dyes.

IMPORTANCE The pulp and paper industry is a major industry in many developing countries, and the global market of pulp and paper wastewater treatment is expected to increase by 60% between 2012 and 2020. Such wastewater contains large amounts of refractory contaminants, such as lignin, whose reclamation is considered economically crucial and environmentally friendly. Furthermore, azo dyes are usually added in order to fabricate anticounterfeiting paper, which further increases the complexity of the pulp and paper wastewater. This work may offer a better understanding of biohydrogen production from xylose in the presence of azo dyes and provide a promising energy-recycling method for treating pulp and paper wastewater, especially for those containing azo dyes.

INTRODUCTION

Pulp and paper wastewater contains large amounts of refractory contaminants, and the pulp and paper industry is a major industry in many developing countries (1–3). The global market of pulp and paper wastewater treatment is expected to increase by 60% between 2012 and 2020 (3). Pulp and paper wastewater contains large amounts of renewable resources, e.g., lignocellulosic wastes, whose reclamation is considered economically crucial and environmentally friendly (4, 5). Coagulation flocculation can be used to recycle the biomass efficiently from liquid, but it is difficult to purify the biomass from the floc, which contains complex components. This may increase the amount of residue sludge and the cost of posttreatment (6–8). Compared to the physical and chemical methods, anaerobic treatment may accomplish the chemical oxygen demand (COD) removal and resolve the economic issue simultaneously in treating pulp and paper wastewater (3, 9–11). Therefore, energy conversion through the anaerobic digestion process has the potential to reclaim methane and/or hydrogen gas from the lignocellulosic compounds and their hydrolysis products, such as carbohydrates, proteins, and fats.

Biohydrogen production from the pulp and paper effluent containing rich cellulosic material could be achieved through both dark and light fermentation processes (12, 13). One of the cellulose hydrolysates, i.e., glucose, could be easily fermented to hydrogen gas (14–16). However, the other abundant hemicellulose hydrolysis product, i.e., xylose, was less efficient as a precursor for biohydrogen production both in pure and in mixed cultures (17–19). A pure culture of a Clostridium beijerinckii strain which follows a butyrate/acetate fermentation pathway was found to ferment xylose to produce hydrogen (20, 21), while other robust biohydrogen producers, e.g., Klebsiella oxytoca, which follows a formate cleavage pathway, were scarcely reported (22). Furthermore, azo dyes represent more than half of the dyes applied in the pulp and textile industries. Some hydrophilic groups contained in azo dyes make them soluble. Such azo dyes are usually added to fabricate anticounterfeiting paper, which further increases the complexity of the wastewater (23, 24).

The objective of this work was to evaluate the feasibility of biohydrogen production from xylose and the effects of azo dyes on biohydrogen performance by Klebsiella oxytoca GS-4-08. Two typical azo dyes with sulfonate and carboxyl groups, methyl orange (MO) and methyl red (MR), were used to evaluate their effects on the strain behaviors and xylose utilizations. Simultaneous decolorization and biohydrogen production processes and their interactions are discussed as well. This work may offer a better understanding of biohydrogen production from xylose in the presence of azo dyes and provide a promising energy-recycling method for treating pulp and paper wastewater.

RESULTS

Decolorization and biohydrogen production by resting cells with xylose as a sole substrate.

The process of xylose degradation and MO decolorization by resting cells in a typical run is illustrated in Fig. 1. The initial xylose, MO, and dry weight (DW) concentrations were kept at 2 g liter⁻¹, 32.7 mg liter⁻¹, and 24 mg liter⁻¹, respectively. The biohydrogen production from xylose could be well described by a modified Gompertz equation (Fig. 1a), and the calculated maximum hydrogen production (P_max,hydrogen), lag time (λ), and maximum hydrogen formation rate (R_max,hydrogen) values were 75.00 ml, 12.64 h, and 2.47 ml h⁻¹, respectively (R² = 0.995). At the end of the batch experiment, a high xylose utilization rate (UR_xyl), 93.99%, was obtained, and the calculated hydrogen molar yield (HMY) was determined as 2.30 mol of H₂ mol of xylose⁻¹. The ability of Klebsiella oxytoca to produce hydrogen by using xylose as the substrate was comparable to that in pure culture or mixed culture (19, 25).

FIG 1.

Profiles of hydrogen production and production rate (a), xylose utilization and fermentation product formation (b), and decolorization of methyl orange (MO) (c). The initial xylose and MO concentrations in the batch experiments were 2 g liter⁻¹ and 32.7 mg liter⁻¹, respectively. All the experiments were performed in triplicate, and data points are the means of triplicate analyses.

According to the degradation pathway of Klebsiella pneumoniae proposed by Chen et al. (26), two main liquid products, i.e., ethanol and acetate, could be produced. These two products were detected in this study and showed similar accumulation trends as biohydrogen production. The final concentrations of ethanol and acetate were 590 and 486 mg liter⁻¹, respectively (Fig. 1b). Moreover, a decolorization efficiency of 83.8% was obtained at 24 h with an initial MO concentration of 32.7 mg liter⁻¹ (Fig. 1c). The results indicate that the xylose can be used as the sole carbon source for simultaneous decolorization and biohydrogen production by K. oxytoca.

Influence of initial xylose concentrations on xylose utilization, main fermentation products, hydrogen production, and decolorization by resting cells.

As shown in Table 1, increasing the xylose concentrations from 2 g liter⁻¹ to 8 g liter⁻¹ obviously reduced the UR_xyl from 93.99% to 65.59%. Similar to the case in some previous studies, a higher UR_xyl was more easily achieved when a lower initial substrate concentration was provided (25). As shown in Fig. 2a, compared to the sharp decrease of UR_xyl, the cumulative hydrogen volumes were raised from 72.7 ml to 104.8 ml with the increase in initial xylose concentrations. However, the calculated HMYs were decreased from 2.30 to 1.19 mol of H₂ mol of xylose⁻¹, and so were the cumulative acetate production/xylose consumed (Y_HAc) and cumulative ethanol production/xylose consumed (Y_EtOH) (Table 1).

TABLE 1.

Effect of initial xylose concentration on the degradation efficiency of xylose and biohydrogen, acetate, and ethanol production by K. oxytoca^a

Cxyl	URxyl	HMY	YHAc	YEtOH
2	93.99 ± 3.49	2.30 ± 0.05	0.909 ± 0.021	1.024 ± 0.049
4	66.44 ± 3.14	2.09 ± 0.09	0.559 ± 0.015	0.903 ± 0.037
6	58.61 ± 4.57	1.70 ± 0.02	0.405 ± 0.013	0.706 ± 0.051
8	65.59 ± 3.26	1.19 ± 0.08	0.333 ± 0.018	0.510 ± 0.027

^a

C_xyl is the initial xylose concentration (grams/liter). UR_xyl is the utilization rate at incubation time of 72 h and was calculated as xylose consumed/initial xylose concentration (percent). HMY (mole per mole) was calculated as:

$$\mathrm{HMY}=(\frac{\text{produced\hspace{0.17em}}{\text{H}}_2\left(\text{ml}\right)}{22.4\text{\hspace{0.17em}ml}/\text{mmol}}\times \frac{273.15\hspace{0.17em}\text{K}}{308.15\hspace{0.17em}\text{K}})\hspace{0.17em}/\text{\hspace{0.17em}xylose\hspace{0.17em}}\mathrm{consumed}\hspace{0.17em}\left(\text{mol}\right)$$

Y_HAc was calculated as the cumulative acetate production/xylose consumed (mole per mole); Y_EtOH was calculated as the cumulative ethanol production/xylose consumed (mole per mole). Results are the means of triplicate experiments.

FIG 2.

Biohydrogen production (a) and decolorization kinetics (b) with different initial xylose concentrations by resting cells. Results are the means of triplicate analyses, and error bars indicate the standard deviations.

A higher initial xylose concentration also enhanced the decolorization process (Fig. 2b), and the pseudo-first-order kinetic rates (k₁) of decolorization were slightly increased, from 0.053 to 0.076 h⁻¹. The decolorization efficiencies were 83.8%, 91.6%, 96.2%, and 98.1% when the initial xylose concentrations were 2, 4, 6, and 8 g liter⁻¹, respectively. The accelerated decolorization in the presence of more xylose indicates that more generated reducing equivalents could accelerate MO decolorization. A similar phenomenon has been previously reported with sucrose as the electron donor (27). Experiments with initial xylose concentrations higher than 8 g liter⁻¹ were not performed in this study due to the poorer values for UR_xyl and HMY.

Influence of MO concentration on biohydrogen production by resting cells.

The maximum hydrogen gas production, acetate and ethanol production showed little difference irrespective of the MO concentration changes (Table 2). The R² values varied between 0.983 and 0.999, indicating that fermentation can be well simulated by a modified Gompertz equation. In the batches with initial MO concentrations of 0, 0.1, and 0.3 mM, the corresponding HMYs were calculated as 2.14, 1.99, and 2.02 mol of H₂ mol⁻¹ xylose, respectively. The relative standard deviation (RSD) of the hydrogen molar yields in the batches with increased MO concentration was only 3.1%. The RSDs of acetate and ethanol were 3.56% and 0.19%, respectively. Both HMYs and fermentation product yields from xylose were independent of the initial MO concentration, indicating that the fermentation and biohydrogen pathways were not disrupted by MO within the tested concentration range.

TABLE 2.

Effects of initial methyl orange concentration on the biohydrogen, acetate, and ethanol production by K. oxytoca^a

Product	CMO (mM)	μb	Pc	λ (h)	R2
Hydrogen	0	3.16	63.59	12.68	0.999
	0.1	3.36	63.64	12.55	0.999
	0.3	2.90	63.78	11.78	0.997
Acetate	0	1.48	39.80	8.40	0.997
	0.1	1.56	38.51	8.55	0.999
	0.3	1.45	41.97	7.92	0.990
Ethanol	0	1.09	33.72	2.39	0.983
	0.1	1.04	33.62	2.49	0.988
	0.3	1.10	33.77	2.35	0.991

^aC_MO, initial methyl orange concentration.

^bEstimated maximum hydrogen production rate (milliliters per hour) and estimated maximum production rates of acetate and ethanol (milligrams per hour).

^cEstimated maximum hydrogen yield (milliliters) and estimated maximum acetate and ethanol yields (milligrams).

Xylose fermentation in the presence of MO and MR by growing cells.

As shown in Fig. 3a, the utilization of xylose by growing cells was initiated at the beginning of the experiments with and without 0.3 mM MO, and UR_xyl was achieved at 88.4% and 86.6% at 72 h, respectively. As expected, the UR_xyl values for growing cells were lower than those for resting cells. But the xylose could only be utilized within the initial 24 h by growing cells in the presence of 0.3 mM MR, and a final UR_xyl of 44.6% was obtained in the end (Fig. 3a). The curves of hydrogen production in presence and absence of MO were almost the same (Fig. 3b), and the final cumulative hydrogen volumes were 60.9 and 63.5 ml, respectively. However, no hydrogen was produced by growing cells during xylose fermentation in the presence of MR (Fig. 3b). The produced acetate and ethanol in the batch experiments were both much lower than those with MO and without azo dyes (Fig. 3c and d). The result indicates that MO did not influence the biohydrogen production by growing cells, but MR did.

FIG 3.

Xylose utilization (a), hydrogen production (b), acetate production (c), and ethanol production (d) in xylose fermentation in the presence or absence of azo dyes by growing cells. All the experiments were performed in triplicate and initiated with a 1% inoculum. Results are the means of triplicate experiments, and error bars indicate standard deviations. Green circles, blue triangles, and red squares, biohydrogen experiments with methyl orange, with methyl red, and without azo dyes, respectively.

Table 3 summarizes the electron equivalent (e⁻ equiv) balance for the biohydrogen and decolorization experiments by growing cells. Under the uncontrolled pH condition, the pH values of the batches decreased from 7.2 to the range of 4.51 to 4.70. Ethanol was the largest final product, containing 32.70 to 33.78% of the electrons in the initial xylose ($e_{\text{xyl},\text{in}}^{-}=26.64$), and acetate and hydrogen were the second most important electron sinks, in ranges of 18.11% to 18.67% and 18.19% to 18.86%, respectively, except for the batches with 0.3 mM MR. The total electron sinks were less than the given e⁻ equiv from xylose. This may result from the unidentified fermentation products, e.g., 2,3-butanediol, formate, and soluble microbial products. The higher calculated total electron sinks of batch experiments with 0.3 mM MR may be attributed to the lower degradation efficiency of xylose, which resulted in negligible unidentified fermentation products.

TABLE 3.

Fractions of electron sinks with different azo dyes and inoculum size

Conditions	Avg (%) of each electron sink of initial xylose for growing cells					Avg (%) of each electron sink of initial xylose for resting cells
	1% (vol/vol) no azo dyes	1% (vol/vol) 0.3 mM MO	1% (vol/vol) 0.3 mM MR	10% (vol/vol) 0.3 mM MR	100% (vol/vol) 0.3 mM MR	No azo dyes	0.3 mM MO	0.3 mM MR
Final pH	4.62	4.60	4.75	4.73	4.7	4.51	4.62	4.70
Compounds
Acetate	18.67	18.11	5.69	6.91	21.72	19.97	20.57	21.72
Ethanol	29.87	30.36	8.77	7.68	33.59	33.78	32.70	33.59
Res-Xylosea	16.87	13.71	77.88	68.47	13.11	13.26	10.91	13.11
DWb	5.58	4.46	3.81	4.83	NDc	ND	ND	ND
H2	18.86	18.19	0.29	0.13	16.07	18.78	19.02	16.07
Azo dye		0.08	0.45	0.45	0.45		0.42	0.45
Total	89.84	84.92	96.89	88.47	84.95	85.79	83.63	84.95
Error	−10.16	−19.55	−6.92	−16.37	−15.05	−14.21	−16.37	−15.05

^aRes-Xylose, residual xylose.

^bDW, dry weight; its chemical formula is C₅H₇O₂N (8 g of COD biomass = 1 e⁻ equiv biomass).

^cND, not detected.

Effects of inoculum size on biohydrogen production in the presence of MR.

Resting cells (inoculum size = 100%) were more tolerant to MR than growing cells; 88.0% of xylose was consumed in the batch experiment with 0.3 mM MR, but only 29.6% and 38.2% were utilized by 1% and 10% inoculum cells, respectively (Fig. 4a). The obtained UR_xyl in MR decolorization by resting cells was comparable to that in MO decolorization (93.99%). As shown in Fig. 4b, c, and d, the biohydrogen, acetate, and ethanol productions in the batch with resting cells were not significantly influenced as well. For the growing cells, the UR_xyl and the production of hydrogen gas and fermentation products were not satisfactory in batch experiments with inoculum sizes of 1% and 10%.

FIG 4.

Xylose utilization (a), hydrogen production (b), acetate production (c), and ethanol production (d) in xylose fermentation in the presence of 0.3 mM methyl red by growing and/or resting cells. Results are the means of triplicate experiments, and error bars indicates standard deviations. Blue triangles, green squares, and red circles, biohydrogen experiments with inoculum sizes of 1%, 10%, and 100% (resting cells), respectively.

As shown in Table 3, ethanol was the dominant end product in the batch with a 100% inoculum size, containing 33.59% of the pregiven electrons (the theoretical production of electrons from full degradation of xylose), which is similar to those without azo dyes or with 0.3 mM MO. Fewer electron sinks of acetate, ethanol, and hydrogen were obtained in the batches with smaller inoculum size, further indicating that the xylose fermentation and biohydrogen production were interrupted by MR. It should be noted that the MR in all batches was completely decolorized in the biohydrogen production period, with the calculated electron sinks of 0.45. The small electron sinks responsible for azo dye decolorization indicate that the electron competition is not the main reason for the xylose fermentation and biohydrogen production.

Biohydrogen production from industrial pulp and paper wastewater.

The feasibility of biohydrogen production from industrial pulp and paper wastewater by K. oxytoca was investigated. Unfortunately, the cell growth in the wastewater was ignorable; the COD degradation efficiency and HMY were only 17.45% ± 2.32% and 4.5 mmol of H₂ per g of consumed COD, respectively. The low concentration of xylose in the collected wastewater might responsible for the low HMY, and it is also insufficient to support the cell growth and to produce economical amounts of hydrogen gas. To confirm this, extra xylose (final concentration was 4 g liter⁻¹) was added to the collected wastewater, and 1.34 ± 0.07 mol/mol of HMY was obtained (Table 4). Although the biohydrogen production from industrial pulp and paper wastewater was lower than that from BDM (Table 2), the feasibility of biohydrogen production was proved if xylose was added.

TABLE 4.

Biohydrogen production from industrial pulp and paper wastewater with extra xylose by resting cells in the presence of toxic compounds^a

Toxic compoundb	URxyl	HMY	YHAc	YEtOH
None	83.67 ± 1.60	1.34 ± 0.07	0.312 ± 0.007	0.540 ± 0.034
Chlorophenol	78.16 ± 1.13	1.22 ± 0.08	0.273 ± 0.005	0.433 ± 0.031
Phenol	84.53 ± 0.79	1.26 ± 0.14	0.324 ± 0.018	0.471 ± 0.024
SO42−	82.63 ± 0.68	1.33 ± 0.09	0.306 ± 0.013	0.494 ± 0.009
Kraft lignin	85.38 ± 0.67	1.25 ± 0.08	0.314 ± 0.019	0.522 ± 0.042

^aUR_xyl (percent), HMY (mole per mole), Y_HAc (mole per mole), and Y_EtOH (mole per mole) were calculated with the same method as shown for Table 2. Results are the means of triplicate experiments.

^bThe concentrations of chlorophenol, phenol, SO₄²⁻, and kraft lignin were 20 mg liter⁻¹, 500 mg liter⁻¹, 2 g liter⁻¹, and 2.5 g liter⁻¹, respectively.

Meanwhile, some possible toxic compounds, which usually exist in industrial pulp and paper wastewater (3, 10), were separately used to evaluate their effects on biohydrogen production performance. As shown in Table 4, it was found that slight effects on biohydrogen production were produced in the presence of 500 mg liter⁻¹ of phenol, 2 g liter⁻¹ of SO₄²⁻, and 2.5 g liter⁻¹ of kraft lignin. However, obvious inhibitions of UR_Xyl, HMY, and yields of fermentation products were produced with only the presence of 20 mg liter⁻¹ of chlorophenol. The high toxicity of chlorophenol to microorganism may result in the lower biohydrogen production, which is in accordance with the reported IC₅₀s (concentrations at which 50% of the activity of strain is inhibited), which range between 0.9 and 76 mg liter⁻¹ (3).

DISCUSSION

This study reports that decolorization and biohydrogen production by K. oxytoca could be achieved simultaneously with xylose as the substrate. But the reduced products of MO, i.e., sulfanilic acid and N,N-dimethyl-p-phenylenediamine (DMPD), could not be further degraded during xylose fermentation. These compounds have been proved recalcitrant to both the pure strain and mixed microorganisms under anaerobic conditions (27–30). The formation of ethanol, acetate, and hydrogen gas during xylose fermentation suggests that the strain was able to ferment xylose to 3-phosphate-glyceraldehyde via the pentose phosphate pathway (PPP), glycolysis, and biohydrogen pathway in turn, which further proves that the strain belongs to the ethanol/acetate formation type (22). The so-called ethanol/acetate fermentation type follows the formate cleavage step by formate-hydrogen lyase, and this step is generally believed to be the dominant mechanism for biohydrogen production of fecal coliform strains of genera such as Klebsiella, Escherichia, and Enterobacter (26, 31, 32). A maximum HMY of 2.30 mol of H₂ mol of xylose⁻¹ was obtained by resting cells in the presence of 2 g liter⁻¹ of xylose in this study, which is higher than the HMY (approximately 2.25 mol of H₂ mol of xylose⁻¹) obtained by the strain that follows the butyrate/acetate pathway without addition of electron shuttles (25).

Moreover, decolorization of MO was achieved during biohydrogen production, indicating that the produced NADH in the acetate pathway would be oxidized by extracellular electron acceptor through the respiratory chain on the cell membrane of the Klebsiella strain (33, 34). That is why the more xylose was provided, the higher the decolorization rate obtained, and this is similar to those six-carbon sugars as the substrate for reduction of azo dyes (28, 35). However, HMY was gradually reduced with increased xylose concentration, and so were the total calculated electron sinks (see Table S1 in the supplemental material). The unfavorable result for hydrogen production in the presence of higher xylose might be the following: (i) most of carbon flow turned to an alternative metabolism pathway such as glycogen synthesis (evidenced by the DW data in Table S1), and therefore less carbon flow went to the hydrogen production pathway (36); (ii) more generated NADH is consumed for fermentation of other by-products, e.g., 2,3-butanediol (evidenced by the data shown in Table S1) (37); and (iii) a hydrogen uptake process exists and is enhanced in this situation; for example, Redwood et al. found that the deletion of the two hydrogen uptake hydrogenases in Escherichia coli strains FTD67 and FTD89 improves the hydrogen production by 50% (38), and Niu et al. reported that the uptake hydrogen of Klebsiella pneumoniae ECU-15 was enhanced with the increases of the initial substrate concentration (39).

Similar to biohydrogen production from disaccharide and hexoses (27), the formations of biohydrogen and liquid fermentation products were not interrupted by sulfonated azo dye-MO when the xylose served as the substrate. In order to explain the specific degradation and metabolism mechanism, we proposed a hypothesis diagram of simultaneous biohydrogen production and decolorization with xylose as the substrate by Klebsiella oxytoca GS-4-08. As shown in Fig. 5, the d-xylose was intracellularly isomerized and phosphorylated to d-xylulose-5-phosphate and then fermented to glyceraldehyde-3-phosphate (Gly-3P) through the PPP (22). When Gly-3P enters into glycolysis pathway, there are two steps capable of generating NADH. One involves the formation of acetate and the other involves transforming H⁺ to H₂ by hydrogenase (26). Under aerobic conditions, those generated NADH would be completely oxidized to synthesize ATP through the respiratory chain on the cytoplasmic membrane (26, 33). However, in this case, under anaerobic conditions as shown in Fig. 5, the electrons may be transferred through the respiratory chain and catalyzed by azoreductase (AzoR) to achieve the unspecific extracellular reduction of MO (40). For growing cells, the biohydrogen production and cell growth were not interrupted by the presence of MO even when the dye concentration was raised to as high as 0.3 mM (Fig. 3b; see also Fig. S1). A possible reason is that the sulfonic azo dye and the reduced products cannot be used as the sole carbon source for cell growth (27) and cannot penetrate the cell membrane and produce inhibition (41). As shown in Fig. S1 and Fig. 3a, the cell growth curves and xylose utilization processes in the absence or presence of MO maintained a high degree of consistency. Although MO could be used as an external electron acceptor, cell growth or xylose utilization could not be significantly stimulated. This may be due to the fact that required electrons for reducing 0.3 mM MO were 0.12 equiv in 100 ml of biohydrogen and decolorization medium (BDM), which accounts for only 0.45% of total electron sinks (Table 3). Therefore, the extracellular reduction of MO was independent of intracellular biochemical reaction, including synthesis of cells, glycolysis, and biohydrogen production.

FIG 5.

Hypothesis diagram of biohydrogen production and decolorization with xylose as the substrate by Klebsiella oxytoca GS-4-08 under anaerobic conditions, based on data presented in references 22, 33, and 26. NADH₂ = NADH + H⁺; Fd = ferredoxin and FdH = reduced form of ferredoxin.

On the other hand, carboxylic azo dye (MR) could penetrate (41) and be reduced intracellularly even under aerobic conditions (33). However, similar to MO, the carboxylic azo dye and the reduced products, i.e., DMPD and o-aminobenzoic acid, could not be used as the carbon source for cell growth as well. As shown in Fig. 3a and Fig. S1, both the xylose utilization and cell growth ceased at the initial 24 h. Meanwhile, the ethanol and acetate started to increase after 12 h but were suspended afterward when xylose fermentation was ceased (Fig. 3c and d). The results indicate that MR partially inhibited the cells from self-growth and stopped producing downstream products through the PPP. According to the electron balance calculation (Table 3), the required electrons for MR reduction account for only 0.45%; thus, the low UR_xyl, HMY, and yields of fermentation products may not have stemmed from the electron competition but were more likely from inhibition of enzymes in the metabolism pathway, such as xylose isomerase and xylulokinase (42). Moreover, additional experiments with glucose and pyruvate as sole carbon sources for growing cells (inoculum size = 1%) were individually performed to evaluate the inhibition sites in the presence of 0.3 mM MR. Similar to the experiments with xylose as the substrate, few liquid fermentation products and hydrogen gas were produced in the presence of MR (data not shown). The ignorable biohydrogen generations from glucose and pyruvate indicate that the enzymes of glycolysis pathway and pyruvate formate lyase were both inhibited by MR (31). To clarify whether the reduced products of MR affect the biohydrogen production, the reduced products of MR, i.e., o-aminobenzoic acid and DMPD, were prepared and added to evaluate their effects on biohydrogen production. The experimental protocols are provided in supplemental material. It was found that the biohydrogen productions in the presence of the reduced products and DMPD were not significantly different from those from positive controls (Fig. S2). Therefore, the results further indicate that MR, and not aromatic amines, inhibited glycolysis and pyruvate fermentation (as indicated in Fig. 5).

As shown in Table 3, the carboxylic azo dye does inhibit biohydrogen production regardless of the inoculum size but slightly influences the synthesis of new biomass. For batch experiments with growing cells, the calculated e⁻ equiv of consumed xylose ($e_{\text{xyl,consumed}}^{-}$) were 5.89 and 8.4, which were higher than the sum of e⁻ equiv of cells and all fermentation products ($e_{\text{products}}^{-}=5.07,5.33$) under 1% and 10% inoculum size conditions, respectively. The e⁻ equiv balance calculations under growing-cell conditions indicated that (i) the generated ATP during xylose consumption is enough for cell growth because synthesis of cells in the presence of MR mainly occurs in the PPP (25, 26) and (ii) similar to the case under aerobic conditions, the toxicity of MR may partially inhibit or delay the synthesis of cells but will not cause cells to die (Fig. S3) during the incubation period (33). Further increasing the inoculum size to 100% (resting cells), we found that the levels of production of hydrogen gas and liquid fermentation products in the batch experiment with 0.3 mM MR were not significantly different from those with 0.3 mM MO and without azo dyes (Table 3). This is probably because the added cells provide more available active sites of enzymes to resist the inhibition from the carboxylic azo dye. In order to confirm this assumption, experiments with inoculum sizes of 30%, 50%, and 70% were conducted to examine the biohydrogen production in the presence of 0.3 mM MR. As shown in Fig. S4a, biohydrogen production from batch experiments with an inoculum size of 30% can be ignored. The calculated electron sinks of products were similar to those from the batch experiments with inoculum sizes of 1% and 10% (Table 3; see also Table S2). When the inoculum sizes were further increased to 50% and 70%, the generated hydrogen gases were similar to that from resting cells, as well as the electron sinks (Table 3).

In summary, the Klebsiella oxytoca strain shows the ability of simultaneous biohydrogen and decolorization with xylose as the substrate. The highest UR_xyl and HMY obtained in this study were 93.99% and 2.30 mol of H₂ mol of xylose⁻¹, respectively, which could compete with those from the butyrate/acetate fermentation pathway. Most importantly, the presented azo dyes were completely reduced during the biohydrogen production period. Although further investigation is needed to identify the enzymatic resistance mechanism when treating wastewater rich in carboxylic azo dyes, we used biohydrogen kinetics and e⁻ equiv balance calculations to postulate that carboxylic azo dye, i.e., MR, penetrates and intracellularly inhibits xylose isomerase, xylulokinase, etc., in the PPP and pyruvate fermentation pathway but that this inhibition effect may be relieved by increasing the initial cell concentration. However, the decolorization of sulfonated azo dye, i.e., MO, was independent of the glycolysis and biohydrogen pathway, most likely due to the extracellular reduction. The feasibility of hydrogen production from pulp and paper effluent by the strain was proved if the xylose is sufficient, and we anticipate that the present finding may offer a promising energy-recycling strategy for treating pulp and paper wastewater, especially that containing azo dyes.

MATERIALS AND METHODS

Dyes and chemicals.

Methyl orange, methyl red, and xylose were purchased from Sigma-Aldrich without further purification. Tryptone, yeast extract, and agar were obtained from Oxoid. N,N-Dimethyl-p-phenylenediamine (DMPD) was purchased from Aladdin. All other chemicals used were of reagent-grade quality or higher, purchased from Sinopharm Chemical Reagent Co., China.

Microorganism and culture conditions.

Klebsiella oxytoca GS-4-08 (CGMCC 5237) was grown aerobically in Luria-Bertani (LB) medium as described previously (33). The biohydrogen and decolorization medium (BDM), unless specified otherwise, consisted of (grams per liter): xylose (2), Na₂HPO₄·12H₂O (1.57), KH₂PO₄ (0.5), (NH₄)₂SO₄ (2), MgSO₄·7H₂O (0.1), and 1 ml of SL-6 trace element solution (33); MO (0.0327 g liter⁻¹) or MR (0.081 g liter⁻¹) was used as the model azo dye. To avoid precipitation of salts during sterilization in an autoclave, MgSO₄·7H₂O, FeSO₄·7H₂O, and SL-6 trace element solution were autoclaved separately. The three solutions were mixed after they were cooled in an aseptic operating board.

Batch experiments.

Klebsiella oxytoca GS-4-08 was first cultured in LB medium at 35°C with shaking at 180 rpm until the end of the exponential phase of growth. After centrifugation (8,000 × g for 15 min) at 4°C, the pellet was washed twice with 50 mM phosphorus buffer solution (PBS; pH = 7.0) and resuspended in PBS to give a final cell density of 2.5 × 10¹⁰ ml⁻¹. Then the cell suspension was transferred to 100 ml of BDM with a predetermined cell concentration for evaluating biohydrogen production and decolorization. For 1% growing cells, 1% (vol/vol) inoculum of the resuspended cell suspension was transferred to 100 ml of BDM, and the final cell concentration was approximately 0.03 ± 0.0018 g (dry weight) liter⁻¹. For 10%, 30%, 50%, and 70% growing cells, the corresponding inoculum sizes were adopted with final cell concentrations of 0.38 ± 0.028, 0.54 ± 0.048, 0.74 ± 0.072, and 0.93 ± 0.089 g (dry weight) liter⁻¹, respectively. For resting cells, all the cells (100% inoculum of the resuspended cells) were transferred to 100 ml of BDM, with a final concentration of 1.17 ± 0.085 g (dry weight) liter⁻¹. To ensure the anaerobic condition, the BDM in serum bottles was purged with high-purity nitrogen gas, and then the bottles were sealed with butyl rubber stoppers and incubated statically at 35°C. All the experiments were performed in triplicate, and results were averaged.

Biohydrogen production from industrial pulp and paper effluent.

The pulp and paper effluent was collected from a sewage conditioning tank at a wastewater treatment plant in a local paper mill factory. The wastewater was stored at 4°C in a refrigerator until use. The concentrations of chemical oxygen demand (COD), Mn, Mg, and Al were determined as 684 ± 28 mg liter⁻¹, 1.588 ± 0.076 mg liter⁻¹, 12.04 ± 0.05 mg liter⁻¹, and 0.55 ± 0.012 mg liter⁻¹, respectively. Fe and Zn were not detected. After flocculation by polyaluminum chloride, the wastewater (COD = 624.5 ± 31.3 mg liter⁻¹) was used for a biohydrogen test. A 1% (vol/vol) inoculum of the resuspended cell suspension (prepared as described above) was transferred to the 100 ml of pretreated wastewater. The same constituents with BDM were also added except for carbon source. The experiments were performed in triplicate, and the average results were used.

Analytical methods.

The liquid samples were collected from BDM at determined time intervals, after centrifugation and passage through a 0.45-μm-pore-size nylon membrane, and the filtrate was stored at 4°C until analysis. Xylose concentrations were determined using the anthrone method (43). MO and MR concentrations were recorded spectrophotometrically (Lambda 25; PerkinElmer, USA) at the dyes' wavelengths of maximum visible absorbance (464 and 436 nm, respectively). COD was determined by a COD digestion vial (5B-1F; Lianhua Inc., China) according to the procedures in the operating manual. The metal concentrations were measured by an inductively coupled plasma atomic emission spectroscopy (ICP-AES) (Optima 2100 DV; PerkinElmer, USA).

Cell numbers, DW, and microbial protein were determined, as well as their relationships, according to reference 33. The viabilities of the growing cells incubated with azo dyes were examined by confocal laser scanning microscopy (CLSM) using a Carl Zeiss LSM710 equipped with a Plan-Apochromat 63× objective. Prior to observation, the cells were stained with a LIVE/DEAD BacLight viability kit (Molecular Probes, Inc.) by following the manufacturer's instructions. The filters used and digital image analysis of CLSM were recommended in reference 44.

Gas production was periodically measured by releasing the pressure in bottles using a glass syringe according to our previous study (27). The hydrogen was measured using a gas-tight syringe (500 μl; SGE Analytical Science, Australia) and analyzed by a gas chromatograph (GC)-thermal conductivity detector (TCD) (GC-9690; FuLi Inc., China). A 5-Å molecular sieve column (100/120 mesh) was used for gas separation, and argon was used as the carrier gas. The two main liquid products, i.e., ethanol and acetate, were detected by another GC (7890A; Agilent, USA) using a flame ionization detector. Operating conditions of the two GCs were according to reference 27.

Kinetic modeling.

To estimate decolorization performance by strain GS-4-08, a modified general kinetic model was used to fit the decolorization curve (28):

$$\frac{dC}{dt}=-{kC}^n$$ (1)

where k is the specific decolorization rate with a unit depending on the values of n (mg^{(1 − n)} dye · 1^{(n − 1)} · h⁻¹); C and n are the azo dye concentration (milligrams per liter) and its partial reaction order, respectively.

A modified Gompertz equation was used to simulate the cumulative products, i.e., hydrogen, acetate, and ethanol production (45):

$$P_i=P_{\max ,i}\exp \hspace{0.17em}\{-\exp \left[\frac{R_{\max ,i}\hspace{0.17em}e}{P_{\max ,i}}(\lambda -t)+1\right]\}$$ (2)

where P_max,hydrogen is maximum hydrogen production, i represents hydrogen, R_max,i is maximum hydrogen formation rate, and λ_i is the lag time.

Calculations.

The e⁻ equiv balance in all batch experiments was calculated by equation 3, which was according to Lee et al. (46), with some modification:

$$e_{\text{xyl},\text{in}}^{-}=e_{\text{xyl,residual}}^{-}+e_{\text{liquid prod}}^{-}+e_{{\text{H}}_2}^{-}+e_{\text{cell}}^{-}$$ (3)

where $e_{\text{xyl},\text{in}}^{-}$ is e⁻ equiv of initial xylose, $e_{\text{xyl},\text{residual}}^{-}$ is e⁻ equiv of residual xylose, $e_{\text{liquid}\hspace{0.17em}\text{prod}}^{-}$ is the e⁻ equiv of liquid products, $e_{{\text{H}}_2}^{-}$ is e⁻ equiv of cumulative hydrogen gas, and $e_{\text{cell}}^{-}$ is e⁻ equiv of synthesized cells at the end of batch tests. Take the calculation of $e_{{\text{H}}_2}^{-}$ for an example:

$$e_{{\text{H}}_2}^{-}=\frac{\text{produced\hspace{0.17em}}{\text{H}}_2\hspace{0.17em}\left(\text{ml}\right)}{22.4\text{\hspace{0.17em}ml/mmol}}\times \frac{273.15\hspace{0.17em}\text{K}}{308.15\hspace{0.17em}\text{K}}\times 2\text{\hspace{0.17em}mequiv\hspace{0.17em}of\hspace{0.17em}}{e}^{-}$$

where K is the temperature in kelvin units and mequiv represents milli-equivalents. All values were experimentally measured.